Muscular strength and body size.

Upper arm strength, right and left grip strength, and several anthropometric measures were recorded during a comprehensive medical examination in an epidemiological study in Tecumseh, Michigan. In the present analysis the relationship between muscular strength and body size was determined to facilitate comparisons of strength among individuals irrespective of differences in size, and more generally to derive sex, age and size specific standards for evaluating results of strength tests. Preliminary regressions of arm strength and summed grip strength on age and twelve size variables were performed. Most of the explained variation in strength variables was accounted for by five size variables, height, weight, biacromial diameter, arm girth, and triceps skinfold thickness. A canonical analysis was performed on the three strength variables and the five selected size variables, age and sex specific. After comparison of the relative weighting of strength variables in the subgroups, the unweighted sum of strength measures was adopted as a strength index. The regressions of the index on the five size variables provide age, sex and size specific means fer use as a standard. Comparison of the multiple correlation coefficients from the regressions with the corresponding canonical correlation coefficients indicates the nearly optimal character of the index. The assessment of physical strength is often of interest in epidemiological investigations. The rationale for including strength measurements is generally to study the relationship of strength to the maintenance of health. Such measurements are used for purposes as varied as aiding in classification by somatotype or inferences as to effects of habitual physical activity. Applications are equally diverse. In a review of studies in cardiovascular disease, to consider only one problem, Weinstein and Epstein (1964) cite prospective, retrospective and cross-sectional studies in occupational, industrial, insurance and military groups, entire communities, and other populations, in all of which strength is assessed. Often only grip strength is used because of the ease of obtaining measurements; however, strength Supported by grants Number HE 14712 and HL 09814 from the National Institutes of Health, U.S. Public Health Service. 2Center for Research in Diseases of the Heart, University of Michigan, Ann Arbor, Michigan 48108. 3School of Health, Physical Education and Recreation, University of Tennessee, Knoxville, Tennessee 37916. Human Biology, February 1976, Vol. 48, No. 1, pp. 147-160. ° Wayne State University Press, 1976 This content downloaded from 157.55.39.35 on Thu, 01 Sep 2016 05:18:16 UTC All use subject to http://about.jstor.org/terms 148 Donald E. Lamphiear and Henry J. Montoye varies among muscle groups (see Clarke 1966, p. 142), so maximum force exerted in other strength tests is sometimes recorded. Strength testing has a long history. Asmussen (1962) noted that strength is correlated with age, especially in growing children. In epidemiological studies, the analysis of strength must therefore be age specific, or the age effect must be removed statistically. It is also well known that large persons, on the average, are capable of exerting more force than small persons. Martin (1918) observed high correlations between muscular strength and weight in boys and girls: .93 and .86 respectively. He expressed strength as a ratio of force exerted in kilograms to body weight in kilograms, still the most common measure for comparing strength in persons of different body size; see Faulkner, et al. (1962). However, Martin (1921) had reservations about the ratio as a measure of strength because (a) mass varies with the cube of length, so strength, he felt, should vary with cube of height, (b) obesity is negatively correlated with the strength-to-weight ratio, and (c) as reported by Martin and Rich (1918), the correlation coefficient between force exerted and body weight, so high in boys, is only .58 in adult males. The validity of the strength-to-weight ratio rests on the assumption that strength is proportional to weight, a relationship that has not been established over a wide age range in either sex. Griffitts (1935) noted that in the ratio of strength to weight, the smaller person has the advantage. It can be seen in Keeney s (1955) study of a group of lean champion weight-lifters that this effect is not just a matter of obesity. The maximum weight lifted increased with body weight, but at a diminishing rate, so that when the ratio of strength to body weight was plotted against body weight, the graph turned down at the higher weights. In most studies, e.g., Clarke (1954, 1957) Tornvall (1963), Rasch and Pierson (1963), Laubach and McConville (1966), and Rogers (1927), measures of muscle size or overall body size, such as arm girth or weight, were more closely related to strength than were linear measurements of the skeleton. Maglischo (1968) found in school girls that strength was more closely related to anthropometric measurements themselves than various indices and ratios of these measurements. Laubach (1969) observed that when weight was held constant, fatness as measured by skinfold thickness was negatively related to strength. In the present analysis, the relationships of arm and grip strength measurements to various anthropometric measurements are considered in both sexes over a broad age range. A simple index of strength is proposed, based in part on a consideration of the results of a canonical correlation analysis. Regression equations are given for the calculation of the comThis content downloaded from 157.55.39.35 on Thu, 01 Sep 2016 05:18:16 UTC All use subject to http://about.jstor.org/terms Muscular Strength and Body Size 149 parison standard from measurements of body size by age range for men and women separately. Expressing observed strength relative to this standard permits the study of variation in strength after the effects attributable to body size have been removed.